ISSN: 2278-4853 Vol 10, Issue 9, September, 2021 Impact Factor: SJIF 2021 = 7.699
Asian Journal of Multidimensional Research (AJMR)
https://www.tarj.in
682
AJMR
5
Action
->
Horror
6
Horror
->
Romance
[Source: Authors]
4.
ANALYSIS
According to the randomly generated data and processing the sameusing an assignment problem,
it is observed that a user prefers watching drama after a comedy series/movie. Moreover, the user
wants to see an action movie/series once they are done watching drama. Horror follows this, and
further, the user shifts to romance. Romance is followed by Scientific fiction, and additionally,
the user chooses to watch thrillers and then back to comedy. These observations show a loop that
the OTT user follows. After analyzing the results, we assume that the user gets bored watching a
particular genre, thus changing the mood as they shift to other genres.
5.
LIMITATIONS
The primary focus of the research paper is to suggest a recommender system that the OTT
platforms could use and develop using a technique of Operations research called Assignment
problem. This paper tries to solve the assignment problem using randomly generated numbers
which could be a limitation to the efficiency of the recommender system. The backbone of this
system is the user's preferences, likings, and viewership. Therefore we cannot accurately
recommend any new series/movie which is trending. Human brains are hard to predict, and so
are their moods. Thus, viewers tend to choose a genre based on their current state of mind.
Hence, this should be taken into account as well.
Moreover, the recommender system data does not consider variables like the user's watch time
and the time taken to complete a series. These details are essential to calculate the probability.
The study fundamentally develops a simple recommender system that could only work
effectively in a given scenario. This system does not assess a change in the user's preference over
time unless the data is collected and the process is repeated.
6.
RECOMMENDATIONS
Since the recommender system is based on the preferences of people, it differs from person to
person.OTT platforms give our individuals personalized suggestions to cut down the
disappointment and the time taken to discover something they would like to watch. However, not
everyone has similar choices. So, in this case, a survey or a method of data collection would have
been a viable option as it would have given us a much better outlook on viewer's choices,
including their favorite genre, instead of relying on randomly generated numbers.
Additionally, acquiring the information about an individual's total time watching a genre would
have made the system more accurate. There is a multitude of variables that would change the
results of the recommender system over time. It is also impossible to always be correct for the
user's motive for watching some specific content. We also need to consider that people change
over time, and their fondness for a particular genre is bound to change with age. Often, children
are more likely to watch animated movies/ series or fictional content because that is what excites
them the most. However, when they grow into teenagers or adults, they start acquiring a taste of
their own and might lose their affection towards a particular genre. Overall, this highlights that
future researchers have an opportunity to improve on this system.
Do'stlaringiz bilan baham: